chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 11 Packages Projects Releases Wiki Activity

Opt tp: tp attn support tp reduce scattered input (#10568)

Browse Source
This commit is contained in:
Yongfei Xu
2025-11-15 18:08:12 +08:00
committed by GitHub
parent 4a10e37ba7
commit d91b16eb16
7 changed files with 275 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
docs/advanced_features/server_arguments.md
View File

@@ -394,6 +394,7 @@ Please consult the documentation below and [server_args.py](https://github.com/s
| `--enable-return-hidden-states` | Enable returning hidden states with responses. | `False` | bool flag (set to enable) |
| `--scheduler-recv-interval` | The interval to poll requests in scheduler. Can be set to >1 to reduce the overhead of this. | `1` | Type: int |
| `--numa-node` | Sets the numa node for the subprocesses. i-th element corresponds to i-th subprocess. | `None` | List[int] |
| `--enable-attn-tp-input-scattered` | Allow input of attention to be scattered when only using tensor parallelism, to reduce the computational load of operations such as qkv latent. | `False` | bool flag (set to enable) |
## Debug tensor dumps
| Argument | Description | Defaults | Options |